INDEX

Explanations

millions and monetary amounts

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rouw

-0.10

 defer

-0.09

led

-0.09

Ã¡d

-0.09

 thousand

-0.09

aida

-0.09

lock

-0.08

 thesis

-0.08

POSITIVE LOGITS

aire

0.21

aires

0.21

naire

0.17

aired

0.14

ths

0.13

naires

0.12

th

0.12

ToOne

0.12

-dollar

0.11

airs

0.11

Activations Density 0.045%